A Novel Evolutionary Algorithm for Scheduling Distributed No-Wait Flow Shop Problems

This study focuses on distributed no-wait permutation flow shop scheduling problems that have many practical engineering backgrounds. The objective is to dispatch jobs optimally to multiple processing centers and ordering them for minimizing the maximum completion time (makespan). First, to solve the problems, a mathematical model is established. Second, a novel evolutionary algorithm is proposed, in which a two-dimensional (2-D) array is designed for solution representation. Based on the problem-specific knowledge, a factory assign strategy and jigsaw puzzle inspired algorithm (JPA) are employed for initializing the population of the evolutionary algorithm. Furthermore, a relative local search is used to improve the performance of the proposed algorithm. Finally, 120 instances with different scales are solved and the results are recorded. Comparisons and discussions show the proposed algorithm has computational competitiveness in solving the concerned problems with makespan criteria.